Intent Classifier
The Intent Classifier is the layer that decides which Theo mode a message belongs to — fast, think, code, image, video, social, learn, research, or others — using regex patterns plus a small AI scorer.
In depth
The classifier owns the mapping between a free-form message and a structured Theo Intent. It runs first for cheap obvious cases (exact verbs, slash commands), then escalates to a reasoning model to score multi-dimensional intent if no clean match is found.
When MCIR is enabled, the classifier feeds a richer pipeline that also blends user memory, behavioral patterns, and recent artifacts before committing to a mode.
Examples
- “draw a flowchart” → flowchart creation
- “summarize this PDF” → notes mode + extract
Also known as
Want to put this concept to work in OpenCharts?
Read moreRelated terms
MCIR (Memory-Completion Intent Resolution)
MCIR is the intent-resolution pipeline that combines fast regex routing, a small reasoning model, behavioral patterns, and memory context to decide which Theo mode and tools should handle each message.
Stop Predicate
A stop predicate is the rule that decides when Theo should end a multi-step generation — e.g. stop after the calendar tool runs, or after N tool calls in research mode.
Tool Call
A tool call is a structured action Theo invokes during a response — like create_flowchart, send_email, or db_run_sql — that produces a real artifact or side effect, not just text.